From d079a28ffc7f21a04983a401f3c8ed8f50e5caa7 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?=E8=B5=B5=E5=BF=A0=E6=9E=97?= <170083662@qq.com> Date: Sat, 28 Feb 2026 11:02:40 +0800 Subject: [PATCH] =?UTF-8?q?feat(ai):=20=E6=B7=BB=E5=8A=A0AI=E6=A8=A1?= =?UTF-8?q?=E5=9D=97=E6=96=87=E6=A1=A3=E5=92=8C=E9=87=8D=E6=9E=84=E5=89=8D?= =?UTF-8?q?=E7=AB=AFAI=E7=BB=84=E4=BB=B6?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - 新增 docs/ai/README.md 包含完整的AI模块配置、建表、API文档 - 重构 src/views/ai/index.vue 组件，移除硬编码BASE_URL和多余参数 - 添加 src/api/ai/backend.ts 统一的AI后端API接口实现 - 集成模型列表、流式对话、非流式对话等功能 - 实现SE流式响应处理和鉴权头自动携带 - 移除历史消息存储和温度参数等冗余功能 --- docs/ai/README.md | 61 +++++++++++++ src/api/ai/backend.ts | 200 +++++++++++++++++++++++++++++++++++++++++ src/views/ai/index.vue | 88 +++++------------- 3 files changed, 281 insertions(+), 68 deletions(-) create mode 100644 docs/ai/README.md create mode 100644 src/api/ai/backend.ts diff --git a/docs/ai/README.md b/docs/ai/README.md new file mode 100644 index 0000000..29bc1b3 --- /dev/null +++ b/docs/ai/README.md @@ -0,0 +1,61 @@ +# AI 模块（Ollama + RAG + 订单分析） + +## 1. 配置 + +见 `src/main/resources/application.yml`： + +- `ai.ollama.base-url`：主地址（例如 `https://ai-api.websoft.top`） +- `ai.ollama.fallback-url`：备用地址（例如 `http://47.119.165.234:11434`） +- `ai.ollama.chat-model`：对话模型（`qwen3.5:cloud`） +- `ai.ollama.embed-model`：向量模型（`qwen3-embedding:4b`） + +## 2. 建表（知识库） + +执行：`docs/ai/ai_kb_tables.sql` + +## 3. API + +说明：所有接口默认需要登录（`@PreAuthorize("isAuthenticated()")`），并且要求能够拿到 `tenantId`（header 或登录用户）。 + +### 3.1 对话 + +- `GET /api/ai/models`：获取 Ollama 模型列表 +- `POST /api/ai/chat`：非流式对话 +- `POST /api/ai/chat/stream`：流式对话（SSE） +- `GET /api/ai/chat/stream?prompt=...`：流式对话（SSE，适配 EventSource） + +请求示例（非流式）： +```json +{ + "prompt": "帮我写一个退款流程说明" +} +``` + +### 3.2 知识库（RAG） + +- `POST /api/ai/kb/upload`：上传文档入库（建议 txt/md/html） +- `POST /api/ai/kb/sync/cms`：同步 CMS 已发布文章到知识库（当前租户） +- `POST /api/ai/kb/query`：仅检索 topK +- `POST /api/ai/kb/ask`：检索 + 生成答案（答案要求引用 chunk_id） + +请求示例（ask）： +```json +{ + "question": "怎么开具发票？", + "topK": 5 +} +``` + +### 3.3 商城订单分析（按租户/按天） + +- `POST /api/ai/analytics/query`：返回按天指标数据 +- `POST /api/ai/analytics/ask`：基于指标数据生成分析结论 + +请求示例（ask）： +```json +{ + "question": "最近30天支付率有没有明显下滑？请给出原因排查建议。", + "startDate": "2026-02-01", + "endDate": "2026-02-27" +} +``` diff --git a/src/api/ai/backend.ts b/src/api/ai/backend.ts new file mode 100644 index 0000000..e7c4413 --- /dev/null +++ b/src/api/ai/backend.ts @@ -0,0 +1,200 @@ +import request from '@/utils/request'; +import type { ApiResult } from '@/api'; +import { API_BASE_URL, TOKEN_HEADER_NAME } from '@/config/setting'; +import { getToken } from '@/utils/token-util'; +import { getHostname, getTenantId } from '@/utils/domain'; +import { getMerchantId } from '@/utils/merchant'; + +export interface AiModel { + name?: string; + id?: string; + [k: string]: unknown; +} + +export interface AiChatRequest { + prompt: string; +} + +function normalizeBaseURL(baseURL: string) { + return baseURL.replace(/\/+$/, ''); +} + +// Keep consistent with `src/utils/request.ts` but usable from `fetch` for SSE. +function getRuntimeApiBaseURL(): string { + try { + const apiUrl = localStorage.getItem('ApiUrl'); + if (apiUrl) return normalizeBaseURL(apiUrl); + } catch { + // ignore + } + return normalizeBaseURL(API_BASE_URL || ''); +} + +function buildAuthHeaders(): Record { + const headers: Record = {}; + const token = getToken(); + if (token) headers[TOKEN_HEADER_NAME] = token; + + // Mirror axios interceptor behavior; add fallbacks to make localhost dev usable. + const tenantId = + getTenantId() ?? (localStorage.getItem('TenantId') || undefined); + if (tenantId) headers.TenantId = String(tenantId); + + const companyId = localStorage.getItem('CompanyId'); + if (companyId) headers.CompanyId = companyId; + + const merchantId = getMerchantId?.(); + if (merchantId) headers.MerchantId = String(merchantId); + + const domain = getHostname?.(); + if (domain) headers.Domain = domain; + + return headers; +} + +function extractText(payload: unknown): string { + if (payload == null) return ''; + if (typeof payload === 'string') return payload; + if (typeof payload === 'number' || typeof payload === 'boolean') { + return String(payload); + } + if (Array.isArray(payload)) return payload.map(extractText).join(''); + if (typeof payload === 'object') { + const obj: any = payload; + const candidates: unknown[] = [ + obj.text, + obj.answer, + obj.content, + obj.data, + obj.message?.content, + obj.delta?.content, + obj.choices?.[0]?.delta?.content, + obj.choices?.[0]?.message?.content + ]; + for (const c of candidates) { + const t = extractText(c); + if (t) return t; + } + } + return ''; +} + +export function normalizeModels(payload: unknown): AiModel[] { + if (!payload) return []; + if (Array.isArray(payload)) { + return payload.map((m) => + typeof m === 'string' ? ({ name: m } as AiModel) : (m as AiModel) + ); + } + if (typeof payload === 'object') { + const obj: any = payload; + if (Array.isArray(obj.models)) { + return obj.models.map((m: any) => + typeof m === 'string' ? ({ name: m } as AiModel) : (m as AiModel) + ); + } + } + return []; +} + +/** + * 后端 AI 模块：GET /api/ai/models + * 前端 baseURL 一般已包含 `/api`，因此这里用相对路径 `/ai/models`。 + */ +export async function aiListModels(): Promise { + const res = await request.get>('/ai/models'); + if (res.data.code === 0) return res.data.data; + return Promise.reject(new Error(res.data.message)); +} + +/** + * 后端 AI 模块：POST /api/ai/chat + */ +export async function aiChat(body: AiChatRequest): Promise { + const res = await request.post>('/ai/chat', body); + if (res.data.code === 0) return extractText(res.data.data); + return Promise.reject(new Error(res.data.message)); +} + +/** + * 后端 AI 模块：POST /api/ai/chat/stream（SSE） + * 用 fetch 以便读取 ReadableStream，并附带鉴权/租户头。 + */ +export async function aiChatStream( + body: AiChatRequest, + opts: { + signal?: AbortSignal; + onDelta: (text: string) => void; + onDone?: () => void; + } +): Promise { + const baseURL = getRuntimeApiBaseURL(); + const url = `${baseURL}/ai/chat/stream`; + + const res = await fetch(url, { + method: 'POST', + headers: { + Accept: 'text/event-stream', + 'Content-Type': 'application/json', + ...buildAuthHeaders() + }, + body: JSON.stringify(body), + signal: opts.signal + }); + + if (!res.ok || !res.body) { + const text = await res.text().catch(() => ''); + throw new Error( + `aiChatStream failed: ${res.status} ${res.statusText}${ + text ? ` - ${text}` : '' + }` + ); + } + + const reader = res.body.getReader(); + const decoder = new TextDecoder('utf-8'); + let buffer = ''; + + while (true) { + const { value, done } = await reader.read(); + if (done) break; + buffer += decoder.decode(value, { stream: true }); + // Some servers use CRLF; normalize to simplify parsing. + buffer = buffer.replace(/\r\n/g, '\n').replace(/\r/g, '\n'); + + // SSE events are separated by blank lines. + let idx = buffer.indexOf('\n\n'); + while (idx !== -1) { + const rawEvent = buffer.slice(0, idx); + buffer = buffer.slice(idx + 2); + idx = buffer.indexOf('\n\n'); + + const dataParts: string[] = []; + for (const line of rawEvent.split('\n')) { + const trimmed = line.trimStart(); + if (!trimmed.startsWith('data:')) continue; + dataParts.push(trimmed.slice(5).trimStart()); + } + if (!dataParts.length) continue; + + const data = dataParts.join('\n').trim(); + if (!data) continue; + if (data === '[DONE]') { + opts.onDone?.(); + return; + } + + let delta = ''; + try { + delta = extractText(JSON.parse(data)); + } catch { + // Some servers stream plain text without JSON. + delta = data; + } + + if (delta) opts.onDelta(delta); + } + } + + opts.onDone?.(); +} diff --git a/src/views/ai/index.vue b/src/views/ai/index.vue index a200cf4..fb5b8b5 100644 --- a/src/views/ai/index.vue +++ b/src/views/ai/index.vue @@ -2,41 +2,26 @@ import { computed, onBeforeUnmount, ref } from 'vue'; import { message } from 'ant-design-vue'; import { - listOllamaModels, - ollamaChat, - ollamaChatStream, - type OllamaChatMessage - } from '@/api/ai/ollama'; - - type Msg = OllamaChatMessage; - - // Hardcode endpoint to avoid going through mp.websoft.top `/proxy`. - // The API methods append `/api/*` paths. - // - // IMPORTANT: do not use `127.0.0.1` in browser production builds: - // it points to the visitor's machine, not your server. - // If you want to use server-local Ollama (`127.0.0.1:11434`), put it behind an HTTPS reverse proxy - // (e.g. `https://ai-api.websoft.top` or same-origin `/proxy`). - const BASE_URL = 'https://ai-api.websoft.top'; + aiChat, + aiChatStream, + aiListModels, + normalizeModels + } from '@/api/ai/backend'; const modelLoading = ref(false); const models = ref>([]); const modelId = ref(''); - const systemPrompt = ref('你是一个有帮助的助手。'); const userPrompt = ref('你好，介绍一下你能做什么。'); - const temperature = ref(0.7); const stream = ref(true); const sending = ref(false); const assistantText = ref(''); const errorText = ref(''); - const history = ref([]); - const canSend = computed(() => { - return !!modelId.value && !!userPrompt.value.trim() && !sending.value; + return !!userPrompt.value.trim() && !sending.value; }); const abortController = ref(null); @@ -51,13 +36,14 @@ modelLoading.value = true; errorText.value = ''; try { - const res = await listOllamaModels({ - baseURL: BASE_URL - }); - models.value = (res.models ?? []).map((m) => ({ - id: m.name, - name: m.name - })); + const res = await aiListModels(); + const ms = normalizeModels(res); + models.value = ms + .map((m) => ({ + id: String(m.name ?? m.id ?? ''), + name: String(m.name ?? m.id ?? '') + })) + .filter((m) => !!m.id); if (!modelId.value && models.value.length) { modelId.value = models.value[0].id; } @@ -70,7 +56,6 @@ }; const clearChat = () => { - history.value = []; assistantText.value = ''; errorText.value = ''; }; @@ -82,27 +67,16 @@ assistantText.value = ''; errorText.value = ''; - const system: Msg = { role: 'system', content: systemPrompt.value.trim() }; - const user: Msg = { role: 'user', content: userPrompt.value.trim() }; - const messages: Msg[] = [ - ...(system.content ? [system] : []), - ...history.value, - user - ]; + const prompt = userPrompt.value.trim(); const controller = new AbortController(); abortController.value = controller; try { if (stream.value) { - await ollamaChatStream( + await aiChatStream( + { prompt }, { - model: modelId.value, - messages, - options: { temperature: temperature.value } - }, - { - baseURL: BASE_URL, signal: controller.signal, onDelta: (t) => { assistantText.value += t; @@ -110,30 +84,15 @@ } ); } else { - const res = await ollamaChat( - { - model: modelId.value, - messages, - options: { temperature: temperature.value } - }, - { - baseURL: BASE_URL, - signal: controller.signal - } - ); - assistantText.value = res.message?.content ?? ''; + // axios 已在拦截器里自动带上 token/tenant + assistantText.value = await aiChat({ prompt }); } - - history.value = [ - ...messages, - { role: 'assistant', content: assistantText.value } - ]; userPrompt.value = ''; } catch (e: any) { // Abort is expected when clicking "Stop". if (e?.name !== 'AbortError') { errorText.value = e?.message ?? String(e); - message.error('请求失败（可能是 CORS 或鉴权问题）'); + message.error('请求失败（可能是后端未启动 / 鉴权 / 租户头缺失）'); } } finally { sending.value = false; @@ -184,13 +143,6 @@ > 流式 - 温度 -